W3C home > Mailing lists > Public > www-forms@w3.org > April 2003

XMLForm Performance Metrics

From: ivelin <ivelin@apache.org>
Date: Thu, 17 Apr 2003 08:21:19 -0500
Message-ID: <017401c304e4$3c935910$0100a8c0@MAUCHI>
To: <XForms@egroups.com>, "freebuilder-xmlform" <freebuilder-xmlform@lists.sourceforge.net>, <www-forms@w3.org>

Accurate performance metrics for web applications is a very complex matter.
The following JMeter test suite is by no means a normative reference, but it
gives an idea of the XMLForm performance. It walks through the Feedback
Wizard demo and touches on most aspects of a real world web application:
processing request parameters, data validation, workflow logic and page
rendering.
When run with 10 concurrent threads on a 1GHz Athlon CPU / 640MB RAM,
against JBoss 2.4/Tomcat 4.0/JDK 1.4 hosted on the same box, it yields the
following numbers:
Without server side XSLT transformation, the average page response time is
100ms.
With server side XSLT transformation, the average page response time is
250ms.

Apparently, utilizing client side XSLT, gains 2.5 times faster response
times. This ratio tends to grow with the complexity of the web pages and
respectively their stylesheets.
You are encouraged to download and try different modifications of the test
suite.

http://www.cocoonhive.org/xmlform/documentation/content/performance.xml

-=Ivelin=-
Received on Thursday, 17 April 2003 09:20:56 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 10 March 2012 06:21:55 GMT